Introduction
Inhaled particulate matter (PM), categorized as PM2.5 and PM1
(aerodynamic equivalent diameters of ≤ 2.5 and ≤ 1 µm,
respectively), has proven to be associated with the increasing
morbidity and mortality from cardiovascular and respiratory diseases (Brauer
et al., 2013; Wang et al., 2014). Primary biological aerosol particles
(PBAPs; about 104–108 cells cm-2) constitute an important
component of PM. They can actively metabolize in the atmosphere with their
mass concentrations ranging from 5.49 to 102 ng m-3 (Zhong et al.,
2016). Furthermore, they play an important role in agriculture, the biosphere,
cloud formation, global climate, and atmospheric dynamics (Brodie et al.,
2007; Despres et al., 2012; Christner et al., 2008; Zhou et al., 2014;
Jaenicke et al., 2005). Fungi, the primary group of PBAPs, include 1.5
million unique species, distributed across rural and urban environments
(Hawksworth et al., 2001). They actively eject their spores with aqueous
jets or droplets into the atmosphere. The global emissions of fungal spores
are estimated as the largest source of bioaerosols (Elbert et al., 2007).
Pioneering studies have reported global fungal emissions to reach 28 Tg per
year and contribute to about 4–13 % of the mass concentration of
PM2.5 (Heald et al., 2009; Womiloju et al., 2003). More recently, some
specific fungal species have been verified to be linked with the occurrence
of public health problems (Morris et al., 2002; Yadav et al., 2004; Bowers
et al., 2012, 2013; Cao et al., 2014; Ryan et al., 2009).
Despite their importance, the abundance, diversity, and community structure
of fungi associated with PM have received limited attention in terms of
research.
Earlier studies on airborne fungal communities, primarily based on culturing
methods, found the dominant phyla to be Ascomycota (AMC) and Basidiomycota
(BMC). Some of the species are considered major pathogens and allergens of
plants, animals, and humans, e.g., Hemileia vastatrix,
Aspergillus, Cryptococcus, and Pneumocystis spp.
(Despres et al., 2012; Smets et al., 2016). While most of the fungal species
remain unknown because cultivable species (typically less than 100) occupy
only a tiny minority of all existing species, advances in nucleic acid
sequencing allow the accurate determination of both cultured and uncultured
microbial communities in environmental samples. For bacterial community
composition, Xu et al. (2017a) investigated the abundance and community of
bacteria in submicron particles during severe haze episodes in Jinan, China.
Later, they discussed the diurnal variation of diverse bacterial communities
in cloud water at Mt. Tai, China (Xu et al., 2017b). For diverse fungi in
Mainz, Germany, Frohlich-Nowoisky et al. (2009) described the fungal
community in coarse (> 3 µm) and fine (≤ 3 µm) PM using internal transcribed spacer (ITS) region
sequencing. Yamamoto et al. (2012) reported the crucial influence of
aerodynamic diameter and season on the fungal taxonomic composition in the
northeastern United States by 454 pyrosequencing. The fungal allergens
clustered in the largest size ranges (> 9 µm) in the
fall season, whereas the pathogens were most abundant in the spring season
and were typically observed in particles with aerodynamic diameters of
< 4.7 µm. Subsequently, DeLeon-Rodriguez et al. (2013)
discussed the effect of tropical storm or hurricane periods on the shift of
airborne fungal species over the upper troposphere. Gou et al. (2016)
described the fungal abundance and taxonomic composition of fungi in PM1
and PM10 in winter in China by 18S rRNA gene sequencing. However, that
study focused on the ambient fungi in total suspended particles (TSP),
PM10, and PM2.5, and was primarily conducted over the ground's
surface; therefore, fungal populations in PM1 at high-elevation sites
were not well accounted for. Diverse microbes at high altitudes (such as in
cloud water and precipitation) can act as nucleating agents for cloud and ice
condensation, influence precipitation patterns (Xu et al., 2017b; Pratt et
al., 2009; Creamean et al., 2013; Bower et al., 2013), and drive
the biogeochemical cycling of elements in ecosystem processes. Hence, it is
essential to advance the knowledge of microbes in PM, especially across the
East Asian regions which are frequently ravished by dust, haze or other
weather phenomenon. During 2013, 2014, and 2015, serious air pollution events
associated with the inadequate use of clean energy in the transport,
domestic, and industrial sectors affected northern China, which includes
several areas with severe air pollution, namely Beijing, Tianjin,
Shijiazhuang, Jinan, and Qingdao. Most researchers focus their attention on the
case study of bacterial abundance and diversity (Gao et al., 2014, 2017a; Xu,
et al., 2017a; Wei et al., 2017). The various physical, chemical, and
biological factors caused by the severe haze or dust episodes may cause shifts in
the bacterial community structure. Moreover, the airborne microbial abundance
and diversity are also effected by seasonal and meteorological factors;
however, the investigations into the seasonal variation of fungal characteristics in
aerosol particles have been very limited.
Mt. Tai (36∘15′ N, 117∘06′ E; 1534 m a.s.l.), the
highest site in the North China Plain, is a tilted fault block mountain, its
height increasing from the north to the south, facing the Japanese islands,
Korean Peninsula, East China Sea, and the Yellow Sea. The vegetation cover is
80 %, with nearly 1000 kinds of plants growing in the area. The number
of tourists, from both China and abroad, visiting this mountain increased
from 5.5 million in 2014 to 5.9 million in 2015. Past investigations in this
region mainly concentrated on the physicochemical characteristics of aerosol
particles and cloud water and their influence on air quality and human
health. Thus far, there have been no studies addressing the diverse fungal
community in aerosol particles at Mt. Tai, necessitating the development of a
reliable knowledge base on the atmospheric aerosols in such scenic
destinations.
The objectives of the present study were: (i) to fill the knowledge gaps
regarding the ambient fungi of PM2.5 and PM1 at a high-elevation
site of East Asia, (ii) to elucidate the size-based differences between the
data of ambient fungal concentration and viable fungal community structure
across different seasons, and (iii) to estimate whether environmental factors
play a role in the variation of fungal characteristics at Mt. Tai.
Sample descriptions and the associated meteorological
characteristics of the atmosphere, including the temperature (T), relative
humidity (RH), PM2.5 mass concentration (MC), PM1 mass
concentration, and fungal cell concentrations on the basis of qPCR analysis
of rRNA copy numbers in
PM2.5 and PM1.
Season
Date of
T
RH
PM2.5
PM1
collection
MC
Fungal SSU
Fungal
Fungal
MC
Fungal SSU
Fungal
Fungal
rRNA gene
spore OC
spore MC
rRNA gene
spore OC
spore MC
copy number
MC
copy number
MC
Unit
∘C
%
µg m-3
104 copy m-3
ng C m-3
µg m-3
µg m-3
104 m-3
ng C m-3
µg m-3
Summer
06/25/15
12.6
98
5.5
11.00
7.15
0.02
BDL
18.00
11.69
0.03
06/26/15
14
97
2.8
3.58
2.33
0.01
BDL
6.51
4.23
0.01
06/27/15
15.1
94.6
52.7
1.01
0.66
0.00
18.1
0.40
0.26
0.00
06/28/15
16.9
84.4
91.1
6.85
4.45
0.01
40.0
3.10
2.02
0.01
06/29/15
17.3
62.6
16.8
4.71
3.06
0.01
13.3
2.79
1.81
0.00
07/03/15
17.7
31.0
15.3
12.20
7.92
0.02
12.7
11.40
7.42
0.02
07/07/15
16.9
84.4
94.0
47.70
31.00
0.08
39.9
4.20
2.73
0.01
07/08/15
17.3
62.6
110.9
22.60
14.71
0.04
42.0
5.31
3.45
0.01
08/07/15
17.4
97.6
13.5
5.64
3.67
0.01
11.4
6.94
4.51
0.01
Autumn
10/22/14
6.7
60.7
40.1
9.34
6.07
0.02
28.1
0.37
0.24
0.00
10/25/14
10.3
80.7
48.1
7.28
4.73
0.01
34.6
45.70
29.73
0.08
10/26/14
11.4
73.6
50.8
6.78
4.41
0.01
31.9
8.26
5.37
0.01
11/03/14
0.3
21.6
4.9
1.57
1.02
0.00
BDL
22.20
14.44
0.04
11/04/14
2.6
33.7
31.6
7.95
5.17
0.01
24.5
6.18
4.01
0.01
11/05/14
4.1
30.9
33.1
3.70
2.41
0.01
25.6
103.00
66.75
0.17
11/06/14
5.1
19.3
22.7
12.70
8.24
0.02
18.0
8.86
5.76
0.01
11/07/14
2.6
34.0
19.8
7.89
5.13
0.01
16.0
6.39
4.15
0.01
11/08/14
2.4
45.7
22.4
14.70
9.54
0.02
17.8
2.92
1.90
0.00
11/09/14
1.1
73.1
77.1
5.56
3.61
0.01
33.5
3.61
2.34
0.01
11/10/14
3.0
49.0
49.2
9.38
6.10
0.02
37.2
16.70
10.87
0.03
11/11/14
2.7
65.4
32.7
27.50
17.85
0.05
25.3
26.30
17.07
0.04
11/12/14
1.0
50.1
51.7
7.50
4.87
0.01
25.7
18.30
11.87
0.03
Winter
12/03/14
-8.9
24.4
13.7
5.03
3.27
0.01
9.7
6.84
4.45
0.01
12/04/14
-11
39.1
35.0
8.68
5.64
0.01
30.6
2.78
1.81
0.00
12/05/14
-10.6
23.4
14.5
1.09
0.71
0.00
13.3
16.20
10.52
0.03
12/06/14
-5.7
11.0
9.1
6.32
4.11
0.01
8.3
4.15
2.70
0.01
12/07/14
-5.4
45.7
38.8
7.90
5.14
0.01
30.9
9.36
6.08
0.02
12/08/14
-7.9
35.7
36.5
1.33
0.86
0.00
29.0
7.17
4.66
0.01
12/09/14
-5.3
16.1
16.5
10.10
6.55
0.02
13.5
5.73
3.72
0.01
12/10/14
-5.6
58.3
9.3
3.24
2.10
0.01
8.1
7.10
4.62
0.01
C – carbon, MC – mass concentration, T – temperature, RH – relative humidity, BDL – below the
detection line.
Materials and methods
Sample collection
At Mt Tai, spring occurs from March to May; summer, June to August; fall,
September to November; and winter, December to February, according to the
environmental temperature. Two middle-volume (100 L min-1) samplers
(TH-150A; Wuhan Tianhong Instruments Co. Ltd., Wuhan, China) were deployed
with particles larger than 2.5 and 1 µm trapped by the impactors
and particles smaller than 2.5 and 1 µm collected on the quartz
filters. The 50 % cutoff aerodynamic diameters are 2.5 and
1 µm. The smaller the aerosol particles, the higher
the collection efficiency. Sixty quartz membrane filters (Pall, NY, USA; 88 mm)
were obtained for 23 h (09:00 to 08:00 the next day) over 8–13 days
during each season from 2014 to 2015 at the summit of Mt. Tai (Table 1). The
blank filters were obtained by placing sterilized quartz microfiber filters
inside the sampler without any operation. Before sampling, all the filters
were baked in a muffle furnace at 500 ∘C for 5 h, placed into
sterilized aluminum foil, and then deposited into a sealed bag. To avoid
contamination, the sampling filter holder and materials used for changing
filters were treated with 75 % ethanol every day. After sampling, the
samples were stored at -80 ∘C until the next analysis. PM2.5
and PM1 mass concentrations were monitored by a synchronized hybrid
ambient real-time particulate monitor (Model 5030; Thermo Fisher Scientific,
Wilmington, DE, USA). Half of the PM2.5 and PM1 filters were used
to analyze water-soluble inorganic ions (NO3-, SO42-,
NH4+, K+, Ca2+, Na+, and Mg2+) by an ambient ion
monitor (URG-9000; URG Corporation, Chapel Hill, NC, USA). The remaining
filters were analyzed in the same batch of laboratory experiments, including
DNA extraction, PCR amplification, quantitative real-time PCR (qPCR), and
Illumina sequencing, except for sample A29 on 9 December 2014 (accidentally
omitted in the first batch of Illumina sequencing). Considering that a part
of the sequences in the two batches of experiments differed, we removed this
sample before quality control. Meteorological data, including relative
humidity, wind speed, wind direction, and temperature, were obtained from
http://www.underground.com at a resolution of 3 h during the sampling
period. The visibility was monitored online by a visibility sensor (Model
PWD22; Vaisala, Finland) with a maximum limit of 20 km.
DNA extraction and PCR amplification
The sample pretreatment and DNA extraction experiments were performed
following a protocol optimized by Jiang et al. (2015). This protocol can
extract sufficient DNA from low-biomass environmental samples (e.g., aerosol
particles) and boosted the DNA extraction efficiency by more than twice as
compared to the non-optimized extraction method. Besides, it has been applied
for studying airborne microbial diversity in different environments (Cao et
al., 2014; Deng et al., 2016; Tong et al., 2017; Gao et al., 2017b). Half of
the filters (about 121.64 cm2 in area) were cut into small
pieces, inserted into 50 mL Falcon tubes that were filled with sterilized
1× PBS buffer, and centrifuged at 200 × g for 3 h at
4 ∘C. The resuspension was collected into a 0.2 µm
Supor 200 PES membrane disc filter. We cut the PES membrane disc filter into
small pieces, heated the pieces to 65 ∘C in PowerBead tubes for
15 min and then vortexed them for 15 min. DNA was extracted according to
the standard PowerSoil DNA isolation protocol (Judd et al., 2016) and
purified by AMPure XP bead purification. A parallel extraction procedure was
performed with the blank filter to check for sample contamination. DNA
concentrations were quantified by a NanoDrop 2000 spectrophotometer
(Thermo Fisher Scientific). The fragments of ITS1 regions were amplified from
genomic DNA by PCR using the forward primer ITS1F
(5′-CTTGGTCATTTAGAGGAAGTAA-3′) and the reverse primer ITS4
(5′-TCCTCCGCTTATTGATATGC-3′), which target the fungal ITS region of the
rRNA gene (Manter et al., 2007). The experiment was conducted using the Gene
Amp® PCR System 9700 (Applied Biosystems, CA,
USA) in a total volume of 50 µL PCR mix containing PCR buffer
(1×), 1.5 µM MgSO4, 0.4 µM of each
deoxynucleotide triphosphate, 0.3 µM each of the forward and
reverse primers, 0.5 U Ex Taq (TaKaRa, Dalian, China), 100 ng template DNA,
and double distilled H2O. The thermal cycling profile was 94 ∘C
for 1 min; 35 cycles of denaturation at 98 ∘C for 20 s, annealing
at 68 ∘C for 30 s, and elongation at 72 ∘C for 45 s; and
final extension at 72 ∘C for 5 min. Three replicates of PCR for
each sample were combined together. The final products were separated by
1.5 % agarose gel electrophoresis and purified using the Qiaquick PCR
purification kit (Qiagen, Valencia, CA, USA). Purified amplicons were
quantified by a Qubit 2.0 fluorometer (Thermo Scientific) and pooled with
equal molar amounts. Sequencing libraries were generated using the Truseq DNA
PCR-Free Sample Prep Kit following manufacturer's instructions. Sequencing
was performed on an Illumina MiSeq instrument (Illumina, San Diego, CA, USA)
with the MiSeq reagent kit V3 (Illumina) according to the standard protocols.
Sequence analyses
After high-throughput sequencing, we removed the chimeric and low-quality
sequences using the FASTX-ToolKit
(http://hannonlab.cshl.edu/fastx_toolkit) and UCHIME algorithm (Edge et
al., 2011) before statistical analysis. The remaining high-quality sequences
were normalized to 7973 reads to compare the different samples effectively.
They were then clustered into operational taxonomic units (OTUs) at a
97 % similarity cutoff using USEARCH software (Version 7.1,
http://drive5.com/uparse/). We used the OTUs as the basis for
estimating the alpha diversity and beta diversity. Alpha diversity
estimators, including Chao1, Simpson's index, and Shannon's index, were
performed by the Quantitative Insights into Microbial Ecology software
(Version 1.8.0, http://qiime.org/scripts/assign_taxonomy.html;
Kuczynski et al., 2011). The taxonomy of ITS sequences was analyzed by RDP
Classifier against the UNITE database (release 7.0,
http://unite.ut.ee/index.php; Koljalg et al., 2013) using a confidence
threshold of 70 %. RDP Classifier was used to determine the taxonomic
composition at the phylum, class, order, family, genus, and species levels
(Koiv et al., 2015; Miettinen et al., 2015). The raw reads were deposited
into the NCBI Sequence Read Archive database under accession number
SRR5146156.
qPCR for ITS regions
To determine the fungal biomass, we performed qPCR (Gao et al., 2017a;
Yamaguchi et al., 2016; Lee et al., 2010) using a CFX96 real-time PCR
detection system (Bio-Rad, Hercules, CA, USA) in 25 µL reaction
mixtures containing 12.5 µL TransStart Green qPCR SuperMix,
1 µL ITS3-KYO2 (5′-GATGAAGAACGYAGYRAA-3′), 1 µL ITS4
(5′-TCCTCCGCTTATTGATATGC-3′), 5 µL sample DNA, and
5.5 µL double-distilled H2O. The amplification followed a
three-step PCR for fungal ITS regions: 40 cycles of denaturation at
95 ∘C for 30 s, primer annealing at 52 ∘C for 30 s, and
extension at 72 ∘C for 30 s. A standard curve was created using
tenfold dilution series of fungal ITS region plasmids. Assuming that the
average fungal genome has about 30–200 rRNA copies, the fungal
concentrations were calculated using the methods described by Lee et
al. (2010) and van Doorn et al. (2007).
Fungal contribution to atmospheric organic carbon
The contributions of fungal spores to organic carbon (OC) were calculated
using mannitol as a biotracer. We assumed 1.7 pg mannitol and 13 pg OC per
spore. To assess the contribution of fungal spores to the OC and to the mass
balance of atmospheric aerosol particles quantitatively, we used the
weighted-average carbon (C) conversion factor of 13 pg C per spore and of 33 pg
fresh weight per spore, which had been obtained earlier as the average carbon
content of spores from airborne fungal species (Bauer et al., 2008; Zhu et
al., 2016; Liang et al., 2017).
Relationships between fungal number concentrations of PM2.5
and PM1 with wind speed and wind direction.
Statistical analyses
To determine the differences in the fungal community variations among
different size fractions, meta-analyses based on the permutation t-test
were conducted using Mothur software (version 1.35.1). The program Metastats
can produce a tab-delimited table to display the mean relative abundance of
the mean, variance, and standard error, together with the p values and
q values. Values were considered significant if p≤0.05 and q≤0.05. The Kruskal–Wallis rank sum test was used to evaluate the seasonal
variation of the microbial community. Boxplots and q values have been
provided for illustration. The relationship between the ambient microbial
community and environmental factors, including PM concentrations and chemical
compositions, was assessed with nonparametric Spearman's rank correlation
coefficients by SPSS 16.0.
Results and discussion
Concentration of fungal spores in PM2.5
and PM1
PM2.5 and PM1 samples were collected during summer, autumn, and
winter at the summit of Mt. Tai. Temporal variations of the mass
concentration and corresponding fungal spore numbers of PM2.5 and
PM1 are summarized in Table 1. PM1 mass concentration was stable
over different seasons, while PM2.5 demonstrated a high seasonal
variation, with higher average concentrations in summer
(44.7 µg m-3) than in autumn (37.2 µg m-3)
and winter (21.7 µg m-3). The values were much lower than
that in the summer of 2006 (123.1 µg m-3; Deng et al., 2011)
and comparable with that in the summer of 2007 (59.3 µg m-3;
Zhou et al., 2009). The average PM1 / PM2.5 ratios were 0.45 in
summer, 0.65 in autumn, and 0.84 in winter, implying that fine particles
dominated in summer, while submicron particles dominated in autumn and
winter.
qPCR revealed an average fungal gene copy number of
9.4 × 104 copies m-3 (ranging from
1.0 × 104 to 4.8 × 105 copies m-3) and
1.3 × 105 copies m-3 (ranging from
3.7 × 103 to 1.0 × 106 copies m-3) in
PM2.5 and PM1, respectively. There is no significant differences
between PM2.5 and PM1 based on the uncertainty estimate
(95 % confidence intervals). Assuming an average rRNA gene copy number of
200 per fungal genome (van Doorn et al., 2007; Lee et al., 2010), we obtained
an average fungal concentration of 467 and 644 spores m-3 in PM2.5
and PM1, respectively. The concentrations at Mt. Tai were lower than
those at surface ground sites, including those in South Korea (ranging
from 9.56 × 101 to 4.2 × 104 cells m-3; Lee
et al., 2010), Austria (1.8 × 104 cells m-3 in urban sites
and 2.3 × 104 cells m-3 in suburban sites; Bauer et al.,
2008), Portugal (ranging from 891 to 964 spores m-3; Oliveira et al.,
2009), and the United States (6450 spores m-3; Tsai et al., 2007). Our
lower values might be ascribed to an underestimation of the fungal numbers.
We used a higher gene copy number of 200 for each microbe studied, whereas
DeLeon-Rodriguez et al. (2013) employed a lower number of rRNA copies of
fungal genomes (30–100 copies per genome). The discrepancy between our
results and those of Lee et al. (2010) might be because of the differences in
sample type, sampling time, and altitude. Lee et al. (2010) focused on the
fungal concentration in TSP by a high-volume TSP sampler
(0.225 m3 min-1) 15 m above the ground in autumn and winter,
whereas we obtained the PM2.5 and PM1 by middle-volume samplers
(0.1 m3 min-1) 1534 m above the ground in summer, autumn, and
winter. It is difficult to explain the disparity between different studies
without uniform guidelines for the sampling and quantitative assessment of
bioaerosols.
Fungal abundance varied seasonally with different size particles in the
near-surface atmosphere. Saari et al. (2015) found that coarse fluorescent
bioaerosol particles (1.5–5 µm) increased in summer, whereas in
winter, these particles primarily existed in smaller particles
(0.5–1.5 µm). The snow cover and decreased biological activity in
winter resulted in the disappearance of microbes from the coarse fluorescent
bioaerosol particles. In this study, the highest fungal concentration in
PM2.5 was observed in summer (641 spores m-3), whereas the highest
value in PM1 was found in autumn (1033 spores m-3), indicating
different origins of fungal spores. Huffman et al. (2010) found that
long-range transport of aerosols and anthropogenic sources such as combustion
influence the fluorescent biological aerosol particles having diameters less
than 1 µm. During the autumn sampling, no obvious straw combustion
phenomena occurred, and we detected some long-range transportation events in
November 2014. Long-range transported airborne PM were mainly derived from
the outer Mongolia regions, well-known to be one of the dustiest places in
East Asia (6 November), Siberia (3 and 12 November), and the Taklimakan and
Gobi desert regions (5 November). Influenced by the air movements from the
desert region, the corresponding fungal abundance increased from
6.18 × 104 to 10.3 × 105 copies m-3 (about
16.7-fold). Similarly, the corresponding fungal abundance influenced by air
parcels from the Siberian, and Taklimakan and Gobi desert regions increased
to 22.2 × 104 copies m-3 and
18.3 × 104 copies m-3,
respectively. Hence, we hypothesized that the
long-range transport of air parcels from north China might have contributed
to the fungal enrichment of PM1. In addition, the increased fungal
abundance might be explained by meteorological diversity (Abdel Hameed et
al., 2012). Low wind speed hinders fungal dispersal owing to the accumulation
effect. According to Almaguer et al. (2014), in Cuba, the calm winds coming
from the southwest direction induce the accumulation of fungal spores over
the northern coast of the island. Lin et al. (2000) observed a strongly
negative correlation between wind speeds of < 4 m s-1 and
fungal concentration; the fungal concentration increased as the wind speed
became higher than 5 m s-1 in the Taipei area. In our present study,
the fungal abundance in PM1 showed no obvious increase under breezy
conditions (wind speed < 2 m s-1) mainly from the southern
direction (Fig. 1). When the wind speed was higher than 2 m s-1, the
fungal abundance increased markedly under the influence of westerly winds. As
the westerly wind velocity increased, the fungal concentration increased
slowly. Meanwhile, in PM2.5, the fungal abundance increased with wind
velocities higher than 2 m s-1, mainly from the northwest direction of
the continental areas, where diverse vegetation grows. The phenomenon implies
that westerly and northwesterly winds might highly induce fungal growth and
abundance in PM at Mt. Tai.
Contribution of spores to OC concentrations and PM mass
OC, accounting for 7–80 % of PM mass, constitutes a significant fraction
of atmospheric aerosols (Yu et al., 2004; Ram et al., 2012; Ho et al., 2012).
Ambient fungi are considered a possible source of OC in PMs. Cheng et
al. (2009) estimated the mean fungal OC concentrations in Hong Kong to be
3.7, 6.0, and 9.7 ng m-3, corresponding to 0.1, 1.2, and 0.2 % of
the total OC in PM2.5, PM2.5-10, and PM10, respectively. In
the present study, the range and average concentrations of fungal
contribution to atmospheric OC and mass concentration for PM2.5 and PM1
are listed in Table 1. The daily averaged concentrations of fungal OC in
PM2.5 and PM1 were 6.1 and 8.3 ng C m-3, respectively, with
the respective contributions to PM being 0.067 and 0.096 %, indicating
that airborne fungal spores as a minor source of carbonaceous aerosols cannot
be ignored at Mt. Tai. The fungal contribution to OC obtained at Mt. Tai was
comparable with that observed at an urban site in Hong Kong
(3.7 ng C m-3; Cheng et al., 2009) but lower than that obtained at an
urban site in Austria (117.9 ng C m-3; Bauer et al., 2008) and a
forest site on Hainan Island (147–923 ng C m-3; Zhang et al., 2015).
The discrepancy between the abovementioned studies can be justified by the
difference in particle type studied (TSP, PM10, PM2.5, and
PM1), fungal concentration, spore carbon content, and assessment method
(e.g., sugar alcohol, cultivation, mannitol, and light microscopy). On the
basis of the same conversion factor of 13 pg C spore-1 by mannitol, the
results were much lower than that obtained at an urban site in Beijing
(0.3 ± 0.2 µ C m-3; Liang et al., 2017), implying a
lower fungal concentration at Mt. Tai than that in Beijing. More studies are
needed to better understand the spatial, temporal, and size distributions of
fungal OC contributions to atmospheric particles in urban areas in the North
China Plain.
The relative abundance of the top five orders and two genera in TSP,
PM10, PM2.5, and PM1.
Taxonomic
Common
RASa
RAFb
References
Samplers
Sample
Concentration
level
fungi
type
or abundance
Genera
Alternaria
11.7
6.2
Adhikari et al. (2004)
Andersen sampler (Thermo Andersen, Smyrna, 300082-5211, USA)
TSP
2.6 %
Dannemiller et al. (2014)
High volume PM10 samplers (Ecotech,Knoxfield, VIC, Australia)
PM10
> 1 %
Alghamdi et al. (2014)
PM2.5 samplers (Staplex Air Sampler Division, USA)
PM2.5
2.6 %
Gou et al. (2016)
Low volume air sampler (BGI, USA)
PM1
> 1 %
Aspergillus
2.3
1.9
Cao et al. (2014)
Air samplers (Thermo Electron Corp.,MA, USA)
PM10 and PM2.5
abundant
Gou et al. (2016)
Low volume air sampler (BGI, USA)
PM10 and PM1
abundant
Order
Pleosporales
18.4
45.4
Rittenour et al. (2014)
Buck Bioaire sampler (A.P. Buck, Inc,Orlando, FL, USA)
TSP
46 %
Yan et al. (2016)
Air samplers (Air Metrics, USA, 5 L min-1)
PM10 and PM2.5
29.4 %
Gou et al. (2016)
Low volume air sampler (BGI,USA)
PM10 and PM1
10–15 %
Xylariales
5.0
14.4
Womack et al. (2015)
SKC Biosamplers (BioSampler SKC Inc.)
TSP
abundant
Gou et al. (2016)
Low volume air sampler (BGI, USA)
PM10 and PM1
0–5 %
Eurotiales
4.8
13.3
Yan et al. (2016)
Air samplers (Air Metrics, USA, 5 L min-1)
PM10 and PM2.5
10.6 %
Gou et al. (2016)
Low volume air sampler (BGI, USA)
PM10 and PM1
10–15 %
Capnodiales
4.4
12.5
Yan et al. (2016)
Air samplers (Air Metrics, USA, 5 L min-1)
PM10 and PM2.5
27.96 %
Gou et al. (2016)
Low volume air sampler (BGI, USA)
PM10 and PM1
∼ 25 %
Polyporales
2.5
6.4
Womack et al. (2015)
SKC Biosamplers (BioSampler SKC Inc.)
TSP
abundant
Yan et al. (2016)
Air samplers (Air Metrics, USA, 5 L min-1)
PM10 and PM2.5
3.6 %
Yamamoto et al. (2012)
Eight-stage Andersen sampler (New StarEnvironmental, Roswell, GA, USA)
PM with aerodynamicdiameter is 2.1–3.3,3.3–4.7, 4.7–5.8,5.8–9.0 and > 9.0 µm
abundant
RASa indicates relative abundance
in submicron particles and RAFb indicates relative abundance in fine
particles.
Statistical comparisons of OTUs, and Chao1 and Shannon indices
among summer, autumn, and winter in PM2.5 and PM1.
Taxonomic diversity and composition of ambient fungi
On average, 509 and 475 OTUs were obtained in PM2.5 and PM1,
respectively, which were higher than those obtained in earlier airborne
fungal studies at the ground level in Beijing, China (34–285; Yan et al.,
2016) and Rehovot, Israel (121–178; Dannemiller et al., 2014). The OTUs
associated with PM2.5 in summer, autumn, and winter were higher than
those associated with PM1, implying more diverse fungal spores in
PM2.5. However, the Shannon and Chao1 indices showed different trends in
PM2.5 and PM1 (Fig. 2). The ambient fungi showed the highest
richness and diversity in winter, followed by autumn and summer. Although
PM1 mass concentration dominated in autumn and winter, the corresponding
fungal diversity was lower than that in PM2.5. Similarly, the dominant
PM2.5 mass concentration in summer presented lower diversity than that
in PM1.
In the fungal community, AMC (89.7 %) and BMC (7.0 %) were the
predominant phyla, and they are known to actively discharge spores into the
atmosphere (Fig. 3a). The remaining phyla were Zygomycota (ZMC) and
Glomeromycota. AMC and BMC present a global pattern across continental
(Austria, Arizona, Brazil, and Germany), coastal (Taiwan, Puerto Rico, and
UK), and marine sites (Pacific, Indian, Atlantic, and Southern Ocean)
(Frohlich-Nowoisky et al., 2012). In continental samples, BMC (64 %)
seems to be more abundant than AMC (34 %), whereas in marine sites, AMC
(72 %) is about 2.6 times more abundant than BMC. Herein, the abundance
of AMC was approximately 12.8 times higher than that of BMC. Members of AMC
have single-celled or filamentous vegetative growth forms that are easily
aerosolized, unlike BMC (Womack et al., 2015). Furthermore, 10 classes
belonging to AMC, 10 to BMC, and 1 to ZMC were observed (Fig. 3b). The
preponderant classes belonging to AMC were Dothideomycetes (37.3 %),
Sordariomycetes (15.0 %), and Eurotiomycetes (6.1 %). The dominant
orders in Dothideomycetes included Pleosporales (14.9 %), Capnodiales
(5.3 %), and Botryosphaeriales (1.6 %) (Fig. 3c). Pleosporales has
been reported to include fungi allergenic to local residents (Rittenour et
al., 2014). The values were lower than those reported in Beijing's PM
(Pleosporales: 29.39 % and Capnodiales: 27.96 %) (Yan et al., 2016).
Likewise, the dominant classes in BMC were Agaricomycetes (4.4 %) and
Tremellomycetes (1.5 %), including the orders Polyporales (2.5 %),
Agaricales (1.6 %), and Tremellales (1.2 %). About 291 taxa from the
genus level were determined, including Alternaria,
Glomerella, Zasmidium, Pestalotiopsis,
Aspergillus, and Phyllosticta. The distribution was
discrepant with that at the ground level, wherein Cladosporium
occupied more than 50 % of total fungi, followed by Alternaria,
Didymella, and Khuskia (Oh et al., 2014). The top five orders
(Pleosporales, Xylariales, Eurotiales, Capnodiales, Polyporales) and genera
(Alternaria and Aspergillus) were commonly observed in
suspended aerosol particles (including TSP, PM10, PM2.5, and
PM1) but showed variable relative abundances, as shown in Table 2. We
attribute this disparity to the different sampling approaches, instruments,
and analysis methods. This aspect needs to be probed and studied in depth in
the future.
Relative abundances of fungal communities of PM2.5 and
PM1 at phylum (a), class (b), and order level (c).
Variance analysis of fungal genera based on the Kruskal–Wallis
rank sum test.
Heatmap analysis of the top 64 fungal genera based on Spearman's
rank correlations (*** p < 0.001; ** p < 0.01; * p < 0.05).
Red arrows indicate that the specific fungi varied significantly in
different seasons.
Implication of the allergenic and pathogenic fungi
To date, about 123 fungal genera (mainly belonging to the phylum AMC) have
been identified to be human allergens (Simon-Nobbe et al., 2008). Of the 11
potentially allergy-inducing AMC species and 1 potentially allergy-inducing
BMC species found at Mt. Tai, the 3 most common species were
Aspergillus flavus, Blumeria graminis, and
Saccharomyces cerevisiae. Aspergillus flavus is a common
human pathogen found in air, and it is also a human allergen and mycotoxin
producer (Adhikari et al., 2004). It is associated with invasive
aspergillosis and superficial infections (Hedayati et al., 2007).
Blumeria graminis, found on the surface of plant leaves, causes
powdery mildew on cereal plants (Belanger et al., 2003). Such pathogens and
allergens are expected to be widely spread around the atmospheric environment
in temperate and tropical zones (Vermani et al., 2010). Our results also
revealed that the abundance of potential allergenic and pathogenic fungal
spores in summer were the highest compared to those in autumn and winter.
Clinicians should consider the fungal spores described herein as a possible
cause of human and plant disease under long exposure to airborne
particles throughout the year, especially in the summer season. Furthermore,
the abundance of the abovementioned allergenic and pathogenic fungal spores
in PM1 was about 3.8 times higher than that in PM2.5 in summer,
implying relatively higher health risks for smaller particles. Residents and
even visitors at Mt. Tai should be warned about this phenomenon.
Metastats analysis showing the fungal genera that are significantly
different among PM2.5 and PM1.
Taxa
PM1
PM2.5
p value
q value
mean
SE
variance
mean
SE
variance
Glomerella
10.51984
0.021813
0.013798
22.49025
0.01807
0.009796
0.000999
0.025543
Zasmidium
6.523201
0.011769
0.004017
12.71881
0.012239
0.004494
0.000999
0.025543
Phyllosticta
2.507228
0.004038
0.000473
5.948659
0.004366
0.000572
0.000999
0.025543
Preussia
0.039161
0.000195
1.10E-06
0.009109
6.81E-05
1.39E-07
0.002322
0.042885
Truncatella
0.030579
0.00024
1.67E-06
0.005152
2.44E-05
1.79E-08
0.002784
0.046274
Umbelopsis
0.027549
0.000252
1.84E-06
0.005369
2.55E-05
1.95E-08
0.001669
0.034675
Sebacina
0.021306
0.000196
1.11E-06
0.001261
1.26E-05
4.77E-09
0.000550
0.022857
Cordyceps
0.020939
0.000137
5.45E-07
0.002518
1.75E-05
9.18E-09
0.001392
0.030848
Size distribution and seasonal variation of fungal communities
Both fungal abundance and fungal community show a seasonal trend across
different size fractions (Awad et al., 2013). Yamamoto et al. (2012) observed
that the pathogenic fungi were mainly detected at PM4.7 (PM with
aerodynamic diameter < 4.7 µm), while the allergenic fungi
existed primarily at PM sizes with aerodynamic diameter
> 9 µm. In the present study, a discrepant size
distribution of the fungal community was observed according to the Metastat
analysis by permutation t-tests (Table 3). Glomerella,
Zasmidium, and Phyllosticta were abundantly enriched in
PM2.5, while the abundance of Preussia, Truncatella,
Umbelopsis, Sebacina, and Cordyceps increased in
PM1. The Kruskal–Wallis rank sum test showed that 6 fungal genera had
apparent seasonal variation (Fig. 4). Glomerella and
Zasmidium increased in autumn and decreased as the particle size
increased. Glomerella was widely found on the surface of leaves,
suggesting that leaf senescence is an important source of fungi in PM2.5
in autumn (Wang et al., 2015). Some crucial environmental factors having a
potential influence on fungal release and growth, such as temperature;
NO2; PM10; SO2; CO; relative humidity (Yan et al., 2016);
radiation; vegetation (Moreau et al., 2016); urbanization; and accidental
events, e.g., dust storms (Prospero et al., 2005), rainfall (Zhang et al.,
2015), hurricanes (DeLeon-Rodriguez et al., 2013), and haze (Yan et al.,
2016), have been identified. Herein Spearman's rank coefficient analysis
indicated that Ca2+, a typical water-soluble inorganic ion from dust,
was negatively related to the prevalence of Glomerella and
Zasmidium in autumn (Fig. 5). The increase of Penicillium,
Bullera, and Geosmithia in winter is ascribed to their
sensitivity to low temperature (Sousa et al., 2008; Abdel Hameed et al.,
2012). The results based on Spearman's rank correlation test analysis support
this notion (Fig. 5, p < 0.01). Humidity, another important
factor for fungal release into the atmosphere either by active or passive
modes, is a crucial factor for the variation in fungal spores such as
Lophium (p < 0.01), Cenococcum
(p < 0.05), Tricholoma (p < 0.05), and
Candida (p < 0.05). In summer, no distinct difference
was observed based on the top 40 fungal genera (Fig. 4). However, some trace
fungal genera presented an inverse correlation with temperature
(Coccomyces, p < 0.01; and Dictyosporium,
p < 0.01), humidity (Botryosphaeria,
p < 0.001; Coccomyces, p < 0.01; and
Dictyosporium, p < 0.01), PM2.5
(Acremonium, p < 0.01; Phoma, p < 0.01), and Ca2+ (Talaromyces, p < 0.01;
Acaromyces, p < 0.01). The crucial environmental factors we
identified contributed to the variation in the fungal community. Due to the
limited culture studies on the mechanism for the effects of environmental
factors on specific fungal spores, the relationship between bioaerosols and
environmental factors still needs to be surveyed over a longer duration.